At 10:15 AM 12/29/01 +0200, Peter L. Peres wrote: >Dwayne, the reference you mentioned cannot be traced down for some reason, >so could you please elaborate a little about what you used in the >practical circuit ? > >Like, what is the expected AC pk-pk level (approximately) on the sensor at >room temperature ? > >And do you tie C and B together on the transistors used for sensors ? > >Peter I had a look for the files on my laptop - I have PIC code but don't have the schematic here - its either on the server back at my shop or may even still exist only in paper form in my note book (also back at my shop). Unfortunately, that is all about 1200 KM away from me right now - I'm at my parents home for the holidays. Based upon memory and what I see in the code, I used a 1 meg resistor from +5V biasing the sensor and switched a 100K resistor in parallel with the 1 meg resistor at about a 40 Hz rate (the backplane rate for my LCD display). I think that the amplifier gain was about 70 - a LM324 is adequate because of the AC coupling. The synchronous detector is 1 section of a 4053 analog switch: the output of the amplifier is a 10 uF cap tied to the armature of the analog switch, when only the 1M resistor is tied to the sensor, the output cap is switched to GND. When the 100K resistor is in parallel with the 1M resistor, the output cap is switched to another 10 uF cap tied to the input of the a/d input. Amplifier gain is set so that the a/d does not over-range while measuring the highest temperature desired but is otherwise not critical. I used the 8 bit a/d convertor of a 12ce673 - but with a dither source and oversampling. I wind up with a noisy 14 bit result that is then heavily filtered for an effective accuracy of 10 or 11 bits. Native 8 bit accuracy is not enough because of the 273C offset between 0K & 0C. As far as the math is concerned, the function is a straight line that passes through absolute zero. The slope is dependant on gain and sensor temperature. All math and calibration is done in degrees Kelvin and converted to degrees C for display. The detail I do not remember for sure is amplifier gain - but its not all that important! Build the circuit, measure the output of the detector. Adjust amp gain so that the a/d does not over-range at the highest desired temperature. My implementation (hardware wise) is almost identical to L Stephen Woodward's design: all 3 sections of a 4053, a LM358 dual op-amp (VCC tied to unreg supply of +17Vdc), something like 8 resistors and 4 capacitors gave me all the processing needed for 2 temperature sensors made of 2n4401 transistors. I wrote an initial pass of the software to prove the concept. Then I shelved it! I'll dig it up when I return and probably post it on Piclist.com. dwayne PS: I think the name of Mr Woodward's article was: Low Cost Precision Thermometry. Try doing a web search on that string. Dwayne Reid Trinity Electronics Systems Ltd Edmonton, AB, CANADA (780) 489-3199 voice (780) 487-6397 fax Celebrating 17 years of Engineering Innovation (1984 - 2001) * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Do NOT send unsolicited commercial email to this email address. This message neither grants consent to receive unsolicited commercial email nor is intended to solicit commercial email. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.